1

Appendix

Exposures to volatile organic compounds in healthcare settings

Preliminary Sampling A preliminary sampling campaign was conducted at a U.S. Veterans Affairs (VA) hospital for five

consecutive days in April of 2009 to develop appropriate sampling protocols and refine the sampling

and analysis methods. VOCs were sampled using active, passive and real-time methods. Each sampling

method has unique advantages and drawbacks for exposure assessment in healthcare facilities. Thermal

desorption tubes used with sampling pumps provide a useful means of personal sampling, the ability to

concentrate low-level chemicals during sampling, and sample stability prior to analysis; however,

adsorbent overload and the need for sampling pumps and tubing (which can be problematic in hospital

settings), limit the applicability of this technique. Evacuated stainless steel canisters equipped with flow

controllers allow whole-air samples to be collected over variable durations, provide large volume

samples for multiple analyses, and do not require sampling pumps.[1] These whole-air samples require

pre-concentration prior to analysis in order to detect low-level VOCs. Limitations of this method are

adsorption onto walls or transformation effects within the sample vessel, which have been addressed in

storage stability studies.[2] Sampling using passive badges is appealing because of the ease of use (no

pumps or tubing are needed) and availability of a variety of sampling badges, but have drawbacks

related to stationary air and the lack of diffusion coefficients for many substances of interest. Real-time

VOC monitors provide immediate measurements of total VOC concentrations, but lack chemical

specificity.

Personal and area VOC sampling was conducted using thermal desorption tubes (CarbotrapTM

300,

Perkin Elmer, Shelton, CT) and 6 liter (L) evacuated canisters, and analyzed using EPA Methods TO-17

and TO-15 respectively.[3, 4] Personal and area samples were also collected for organic vapors using

passive badges (Series 575 Passive Sampler for Organic Vapor, SKC, Inc., Eighty Four, PA) analyzed

2

by OSHA Method 1002[5] with a mass spectrometer detector. The results of the preliminary sampling

(data not shown) indicated elevated concentrations of alcohols amidst lower concentrations of other

VOCs of interest. This combination of chemicals (greater than 50 identified per sample) and

concentrations (ppb to ppm range) presented a challenging mixed exposure environment for sampling

and analysis. The multi-bed sorbent in the thermal desorption tubes was overwhelmed by high-levels of

alcohols, making them unsuitable for quantifying low-level VOC exposures in healthcare settings.

Passive badges were also unsuitable for sampling in this work environment because of a lack of

quantifiable adsorption kinetics as well as poor sensitivity for many low-level analytes. In contrast,

evacuated-canister, whole-air samples offered a uniquely suitable approach for handling this wide range

of exposures because a single sample could be split for multiple analyses at both µg/m3- and mg/m

3-

level concentrations.[2]

Exposure Estimates Geometric means and standard deviations for VOC measurements are provided by occupation and

sampler type in Table S1. In addition to the indoor air quality guidelines mentioned in the main text, the

US Green Building Council promulgates a framework for the testing and certification of green building

design, construction, operation, and maintenance; the framework is termed Leadership in Energy and

Environmental Design (LEED) and includes consensus standards for VOC concentrations.[6] LEED

recommends less than 200 μg/m3 for the sum of VOCs measured by EPA Compendium of Methods for

the Determination of Air Pollutants in Indoor Air,[7] which includes IP-1A (canister-based method) and

IP-1B (sorbent-based method). This value is consistent with the German government’s

recommendations and is two to three times lower than those of Japan and Hong Kong. TVOCMIX, an

underestimate of TVOC, clearly exceeded these stringent recommendations by an order of magnitude or

more. These recommendations were purportedly established to improve/maintain indoor air quality for

health, comfort, and productivity.[8] While the strategy of the Canadian National Research Council for

3

accomplishing these VOC levels, including source contaminant control and dilution ventilation, are

practical and appropriate, these levels may not be realistically achieved in a hospital setting due to the

variety of sources present and the general absence of fixed location sources. Hand sanitation and

cleaning product use, for example, are two primary sources of VOC exposure in healthcare settings.

These exposures are difficult to control at the source due to their ubiquitous use throughout a facility.

Qualitatively identified compounds are displayed in Table S2.

4

Table S1 Geometric mean (GM) in µg/m3 and geometric standard deviation (GSD) for area and personal samples by occupation

Occupation Area or

Personal

TVOCMIX ethanol 2-propanol benzene toluene ethylbenzene m,p-xylene o-xylene

GM GSD GM GSD GM GSD GM GSD GM GSD GM GSD GM GSD GM GSD

NA A 4179 1.8 1909 3.1 1650 1.5 0.73 1.9 4.0 1.7 0.35 2.3 1.4 2.1 0.38 1.9

P 9153 1.7 4899 2.6 2936 1.6 2.2 2.0 9.9 2.9 1.2 15 5.4 2.8 1.5 17

LPN A 6686 1.5 4591 1.8 1910 1.4 1.10 2.5 6.9 1.9 1.4 5.0 3.8 3.5 1.4 5.0

P 8680 2.6 3708 7.0 1349 3.8 4.2 5.0 28.8 4.3 2.1 8.7 8.0 3.7 2.6 8.0

MEP A 2605 3.6 32.5 - 1971 5.0 0.42 2.3 2.3 3.9 0.55 6.7 1.8 4.5 0.58 6.3

P 7895 7.6 2.7 - 4566 11 0.15 - 7.5 3.2 <0.16 -* 0.35 19 <0.19 -*

RT A 1361 2.2 759 2.7 160 - 0.73 2.0 17.3 3.3 1.2 4.2 3.3 2.6 1.2 4.0

P 4640 3.3 984 11 70.4 - 0.64 2.7 28.9 6.9 0.51 7.4 4.0 3.4 1.0 5.6

PT A 4315 3.0 115 - 365 - 1.1 1.7 11.1 2.4 2.3 6.2 12.7 3.9 1.8 4.6

P 4123 5.9 176 19 290 - 1.2 1.6 5.5 5.6 <0.16 -* 1.2 4.4 0.45 5.3

FSW A 2682 3.9 601 3.9 1607 4.5 0.68 2.7 3.8 1.7 1.5 4.9 6.2 4.7 2.1 3.7

P 3915 3.6 1308 3.2 1859 5.1 0.88 3.8 20.7 4.0 1.4 5.4 7.7 2.4 1.3 4.7

RN A 2808 2.9 685 14 533 - 0.69 2.0 4.0 3.7 0.37 7.2 1.2 5.8 0.60 6.5

P 3849 2.8 1785 3.9 778 12 1.1 2.5 5.1 2.9 <0.16 -* 0.39 13 <0.19 -*

DA A 9336 1.4 <0.13 -* 9014 1.4 0.88 3.1 11.5 2.1 0.89 11 5.7 3.9 1.2 9.2

P 3755 4.5 0.41 -* 2696 6.9 <0.083 -* 28.7 2.0 <0.16 -* <0.18 -* <0.19 -*

HK A 2990 2.2 1119 2.9 1437 2.7 0.50 2.8 4.0 3.9 0.44 6.2 1.9 4.2 0.45 5.8

P 3254 1.9 1363 2.0 1123 3.4 0.70 4.1 23.6 4.5 0.50 7.9 1.9 7.2 0.73 6.3

ST A 1731 2.3 446 1.4 1129 2.9 0.45 1.4 2.0 1.0 <0.16 -* 0.39 1.1 <0.19 -*

P 2423 2.3 1031 2.7 1077 2.5 0.98 1.2 112 1.3 <0.16 -* 1.8 1.3 <0.19 -*

ET A 1823 5.8 129 - 789 -* 0.59 3.4 4.9 3.2 0.45 11 1.9 4.5 0.33 12

P 1731 4.8 327 7.1 777 7.1 0.17 18 9.8 7.0 <0.16 -* 0.63 - <0.19 -*

CLT A 876 3.5 157 16 66.8 16 0.68 2.4 10.3 1.7 4.1 15 14.1 17 3.9 13

P 1615 4.6 431 3.6 203 1.8 <0.083 -* 162 2.5 2.7 - 25.6 - 2.9 -

MAT A 1034 1.4 336 1.1 532 1.5 0.74 1.3 8.6 1.7 <0.16 -* 0.75 1.2 <0.19 -*

P 1010 1.0 272 1.0 435 1.2 0.83 1.4 137 1.5 4.7 1.3 28.6 1.3 7.0 1.3

DLT A 1720 1.2 279 1.1 1371 1.3 0.93 2.9 10.6 3.4 0.23 -* 2.0 6.5 <0.19 -*

P 490 2.4 198 3.0 132 3.8 <0.083 -* 8.8 2.7 <0.16 -* <0.18 -* <0.19 -*

FB A 2.0 4.7 <0.13 -* <0.14 -* <0.083 -* <0.094 -* <0.16 -* <0.18 -* <0.19 -*

P 40.0 7.2 3.3 -* 3.3 17* <0.083 -* 0.20 17 <0.16 -* <0.18 -* <0.19 -*

Outside A 107.0 1.9 23.8 4.4 33.6 5.0 0.48 2.7 3.1 3.9 0.25 8.4 1.2 5.1 0.31 7.0

Note: dash (-) = GSD > 20 or GM < LOD; asterisk (*) = ≥50% of values below LOD. Occupation Codes: NA=Nursing Assistants; LPN=Licensed Practical Nurses; MEP=Medical

Equipment Preparers; RT=Respiratory Therapists; PT=Pharmacists/Pharmacy Technicians; FSW=Floor Strippers/Waxers; RN=Registered Nurses; DA=Dental Assistants;

HK=Housekeepers; ST=Surgical Technologists; ET=Endoscopy Technicians; CLT=Clinical Laboratory Technicians; MAT=Medical Appliance Technicians; DLT=Dental

Laboratory Technicians; and FB=Field Blanks.

5

Table S1 (continued) Geometric mean (GM) in µg/m3 and geometric standard deviation (GSD) for area and personal samples by

occupation

Occupation Area or

Personal

acetone hexane methyl methacrylate methylene chloride chloroform α-pinene d-limonene

GM GSD GM GSD GM GSD GM GSD GM GSD GM GSD GM GSD

NA A 57.7 1.7 0.40 2.6 <0.17 -* 0.52 2.0* 0.64 2.3 <0.28 -* 6.1 2.7

P 105 1.6 1.4 3.3 1.2 12 2.6 3.0 0.79 8.4 0.43 13* 7.4 4.6

LPN A 63.7 1.4 0.63 3.3 <0.17 -* <0.36 -* 2.5 3.1 0.63 2.2 13 2.3

P 139 3.6 1.8 19 <0.17 -* 3.1 7.1 2.1 7.0 <0.28 -* 3.9 7.0

MEP A 23.8 1.2 0.84 1.7 <0.17 -* <0.36 -* 0.77 4.4 <0.28 -* 7.8 2.2

P 97.8 4.3 2.0 8.0 0.48 - <0.36 -* <0.14 -* <0.28 -* 12.9 -

RT A 29.7 1.4 0.85 2.2 <0.17 -* 0.72 3.6 1.3 3.3 0.49 4.3 4.2 1.3

P 99 5.0 2.2 1.6 <0.17 -* 4.5 1.9 <0.14 -* <0.28 -* <0.53 -*

PT A 59.9 1.2 1.7 3.1 0.38 3.1 0.72 5.3 0.63 2.4 0.81 2.3 10.6 2.3

P 66.3 2.5 3.9 1.9 <0.17 -* 3.8 1.6 <0.14 -* <0.28 -* 1.2 5.0

FSW A 75.1 3.1 0.81 3.1 0.27 3.1 <0.36 - 0.80 1.6 0.71 3.7 12.1 6.6

P 96.1 1.5 1.0 3.7 <0.17 -* 0.63 10 0.35 5.3 <0.28 -* 1.9 -

RN A 44.5 2.3 1.1 2.9 <0.17 -* <0.36 -* 0.91 3.1 <0.28 -* 5.2 2.9

P 73 1.5 0.69 9.1 <0.17 -* 0.94 5.9 0.23 9.9 <0.28 -* <0.53 -*

DA A 42.5 1.5 0.56 1.4 36.5 2.5 <0.36 -* 1.4 5.2 <0.28 -* 5.6 1.4

P 54.6 1.6 0.37 3.3 2.1 14 <0.36 -* <0.14 -* <0.28 -* <0.53 -*

HK A 46.4 1.6 0.57 2.3 <0.17 -* <0.36 -* 1.1 3.3 0.57 2.3 11.3 3.4

P 114 1.5 1.2 1.8 <0.17 -* 0.44 5.6* 0.64 5.8 <0.28 -* 6.1 4.8

ST A 34.9 1.1 0.12 1.9* <0.17 -* <0.36 -* 0.40 2.1 <0.28 -* 2.1 1.4

P 71.7 1.0 0.14 5.2* <0.17 -* 3.1 1.1 0.18 4.0* <0.28 -* 2.8 1.1

ET A 30.4 1.5 0.63 1.6 <0.17 -* <0.36 -* 1.5 3.8 <0.28 -* 2.8 2.8

P 78.5 2.3 0.44 14 <0.17 -* <0.36 -* <0.14 -* <0.28 -* <0.53 -*

CLT A 60.7 6.0 0.30 1.8 <0.17 -* <0.36 -* 0.53 5.1 <0.28 -* 6.3 1.4

P 160 3.0 2.1 2.2 <0.17 -* 4.0 1.5 <0.14 -* <0.28 -* 2.0 3.1

MAT A 71.6 2.1 0.36 2.0 <0.17 -* <0.36 -* 0.38 1.3 0.8 1.1 67.8 1.4

P 78.7 1.1 1.7 1.3 <0.17 -* 2.4 1.1 0.21 11* <0.28 -* 23.6 1.1

DLT A 29.5 1.2 0.30 1.5 10.4 4.0 <0.36 -* 0.61 6.4 <0.28 -* 9.5 1.1

P 43.9 1.9 <0.11 -* <0.17 -* <0.36 -* <0.14 -* <0.28 -* <0.53 -*

FB A <0.19 -* <0.11 -* <0.17 -* <0.36 -* <0.14 -* <0.28 -* <0.53 -*

P 14.7 8.8 <0.11 -* <0.17 -* <0.36 -* <0.14 -* <0.28 -* <0.53 -*

Outside A 16.1 1.6 0.49 2.1 <0.17 -* <0.36 -* 0.21 6.2 <0.28 -* 0.59 8.7*

Note: dash (-) = GSD > 20 or GM < LOD; asterisk (*) = ≥50% of values below LOD. Occupation Codes: NA=Nursing Assistants; LPN=Licensed Practical Nurses; MEP=Medical

Equipment Preparers; RT=Respiratory Therapists; PT=Pharmacists/Pharmacy Technicians; FSW=Floor Strippers/Waxers; RN=Registered Nurses; DA=Dental Assistants;

HK=Housekeepers; ST=Surgical Technologists; ET=Endoscopy Technicians; CLT=Clinical Laboratory Technicians; MAT=Medical Appliance Technicians; DLT=Dental

Laboratory Technicians; and FB=Field Blanks.

6

Table S2 Qualitatively identified compounds by NIST Mass Spectral Library with ≥ 75% quality factor

7

Supporting Task and Product Information An understanding of the tasks performed and products used by occupation (Table S3) are useful

in the interpretation of the principal component analysis below. Tasks performed are listed in the

table with associated occupations who perform the tasks and chemical groups that are contained

in the products used. Products had to be used for at least 15 minutes to be included in the table.

8

Table S3 Occupation and chemical group corresponding to products used for specific tasks

Tasks where Products

Used

Occupation Chemical Groups of Products Used

Bathroom cleaning HK Fragrance

Carboxylic acid

Alcohol

Amide

Amine

Quats

Salt

Acid

Dental laboratory

procedures

DLT, DA, CLT Metal

Acrylate

Oxidizers

Alcohol

Metalloid

Carboxylic acid

Salt

Halogenated

Phenolic

Amine

Ester

Deodorizing HK, DA

Surfactant Alcohol Phenolic

Floor cleaning HK, FSW Surfactant

Fragrance

Glycol ether

Ether

Quats

Acid

Aromatic

Salt

General surface cleaning HK, MEP, RN, ET,

DA, LPN, FSW, RT,

CLT, NA

Alcohol

Quats

Salt

Base

Amine

Ether

Carboxylic acid

Terpene

Oxidizers

Phenolic

Surfactant

Acid

Glycol ether

Amide

Metalloid

Glass cleaning HK, FSW Amine

Carboxylic acid

Ether

Fragrance

Alkane

Glycol ether

Hand cleaning RN, RT, NA, MEP,

ST, SPN, HK, ET

Alcohol

Alkane

Halogenated

Salt

Phenolic

Amide

Instrument cleaning MEP, ET, RN, LPN,

FSW, ST, CLT

Enzyme

Surfactant

Alcohol

Amine

Acid

Amide

Salt

Clinical laboratory

procedures

CLT, LPN, DLT, ET Aldehyde

Salt

Alcohol

Aromatic

Carboxylic acid

Metal salt

Patient cleaning RN, ET, NA, PT,

RT, MEP, LPN, DA,

HK, CLT

Alcohol

Amide

Ether

Amine

Salt

Ester

Glycol ether

Carboxylic acid

Aromatic

Quats

Alkane

Enzyme

Oxidizer

Metalloid

Surfactant

Stainless Steel cleaning

MEP, HK Alkane Ester

Sterilizing & high-level

disinfecting

ET, LPN, MEP Carboxylic acid

Oxidizer

Aldehyde Ether

Stripping and finishing

floors

FSW, HK Glycol ether

Amine

Aromatic

Base

Ether

Ammonia

Ester

Alcohol

Alkane

Acid

Acrylate

Notes: NA=Nursing Assistants; LPN=Licensed Practical Nurses; MEP=Medical Equipment Preparers; RT=Respiratory Therapists;

PT=Pharmacists/Pharmacy Technicians; FSW=Floor Strippers/Waxers; RN=Registered Nurses; DA=Dental Assistants; HK=Housekeepers;

ST=Surgical Technologists; ET=Endoscopy Technicians; CLT=Clinical Laboratory Technicians; MAT=Medical Appliance Technicians;

DLT=Dental Laboratory Technicians; and FB=Field Blanks.

9

Principal Component Analysis Principal component analysis was used to analyze the log-transformed area sample data for the

14 target VOC analytes as inputs; field blanks (n=40) and outside (n=1) samples were excluded

from analysis. Kaiser criterion (eigenvalue > 1) and scree plots indicated that five principal

components captured 68.7% of the variance. Principal components 1 through 5 explained 32.3,

11.9, 8.81, 8.09, and 7.67% of the variance, respectively. These five components were

subsequently rotated using a maximization of the variance (varimax rotation) to produce

orthogonal factor loadings.

The analyte influence on the five factors is displayed as factor loadings in Figure S1A.

Positive values indicate a positive influence on the factor while the converse is true for negative

values. The following analyte influence is apparent from the factor loading distribution as

indicated by open diamonds above the bars when the factor loading was greater than 0.4 or less

than -0.4 (Figure S1a): Factor 1 – benzene, chloroform ethylbenzene, m,p-xylene, o-xylene, and

toluene, which are mostly aromatics and may be indicative of solvent use; Factor 2 – d-limonene

and α-pinene, which are terpenes and may be indicative cleaning product use or due to the

sampling environment from natural sources such as citrus fruit and coniferous trees; Factor 3 –

2-propanol, which is an alcohol used for disinfection; Factor 4 – toluene, which is an aromatic

and may be indicative of solvent use; and Factor 5 – methyl methacrylate, which is a monomer

of acrylic resin used in dentistry. All the chemicals listed in the factors above were positively

correlated with the factor except for methyl methacrylate, which was negatively correlated.

Factor loadings were not influenced by three chemicals: ethanol, acetone, and methylene

chloride. While specific sources could not be associated with each factor, these factors may

represent a combination of tasks, occupations, and chemical groups contained in the products

(Table S3).

10

The factor loadings are subsequently linked to the occupations as mean factor loadings in

Figure S1b. The previous interpretation of positive and negative influence of analyte on factor is

analogous to this interpretation of factor on occupation. This part of the figure may be used to

relate the factor loadings, which are indicative of analyte influence, on the occupations that were

measured to identify trends in exposure among specific occupations. Factor 1 (i.e., solvent use) is

positively correlated with PT and CLT, which is consistent with observations that, e.g., CLT

used aromatics for clinical laboratory procedures, but is negatively correlated with NA, LPN,

MEP, DA, HK, ET, and DLT who more often perform cleaning tasks rather than laboratory tasks

with solvents. Factor 2 is positively correlated with LPN, PT, FSW; limonene and pinene are

fragrances used in floor cleaners and their positive associations in this factor are in line with

product usage by FSW and these compounds are likely to be exposure sources for LPNs and

other occupations that work among various locations in a ward. Factor 2 is negatively correlated

with RT and ET and may reflect the relative immobility of these professions to specific areas

within a ward. Factor 3 is positively correlated with DA which is linked with disinfectant use of

2-propanol by the previous analysis and possibly general cleaning performed by this occupation

between patients. Factor 3 is negatively correlated with RT and CLT, who may not perform these

disinfectant tasks or infrequently use products containing 2-propanol. Factor 4 is positively

correlated with RT and DA; the link between these occupations and toluene is not fully

understood but could, in the case of DA, be related to use of adhesives or glues that contain

toluene but were not captured in our product inventory. Factor 5 is not appreciably positively

correlated with any occupation; factor 5 is negatively correlated with MEP, PT, DA, CLT, and

DLT, but most notably with DA and DLT, which is contrary to expectation. The lack of

association with occupations may be due to the following: the inability of this factor to be

11

realistically interpreted since factor 5 describes the least amount of variance retained in the

model; the relatively small magnitude of the methyl methacrylate exposure compared to alcohols

and aromatics.

12

Figure S1 Factor loadings for by analyte (a) and occupation (b) based on area samples.

Notes: NA=Nursing Assistants; LPN=Licensed Practical Nurses; MEP=Medical Equipment Preparers; RT=Respiratory

Therapists; PT=Pharmacists/Pharmacy Technicians; FSW=Floor Strippers/Waxers; RN=Registered Nurses; DA=Dental

Assistants; HK=Housekeepers; ST=Surgical Technologists; ET=Endoscopy Technicians; CLT=Clinical Laboratory Technicians;

MAT=Medical Appliance Technicians; DLT=Dental Laboratory Technicians; and FB=Field Blanks.

13

References

1 Ochiai N, Daishima S, Cardin D. Long-term measurement of volatile organic compounds

in ambient air by canister-based one-week sampling method. Journal of Environmental

Monitoring 2003;5:997-1003.

2 LeBouf RF, Stefaniak AB, Virji MA. Validation of evacuated canisters for sampling

volatile organic compounds in healthcare settings. Journal of environmental monitoring :

JEM 2012;14:977-83.

3 EPA. Compendium Method TO-17: Determination of Volatile Organic Compounds

(VOCs) in Ambient Air using Active Sampling onto sorbent tubes.

http://www.epa.gov/ttnamti1/files/ambient/airtox/to-17r.pdf. Published January 1999.

Accessed May 2014.

4 EPA. Compendium Method TO-15: Determination of Volatile Organic Compounds

(VOCs) in Air Collected in Specially-Prepared Canisters and Analyzed by Gas

Chromatography/Mass Spectrometry (GC/MS

http://www.epa.gov/ttnamti1/files/ambient/airtox/to-17r.pdf. Published January 1999.

Accessed May 2014.

5 OSHA. Xylenes (o-, m-, p-isomers) Ethylbenzene, 1002. OSHA method.

https://www.osha.gov/dts/sltc/methods/mdt/mdt1002/1002.html. Published August 1999.

Accessed May 2014.

6 USGBC. LEED 2009 for Healthcare.

http://www.usgbc.org/Docs/Archive/General/Docs8878.pdf. Published November 2010.

Updated April 2012. Accessed May 2014.

7 EPA. Compendium of Methods for the Determination of Air Pollutants in Indoor Air.

Environmental Protection Agency report EPA/600/4-90-010.

http://www.epa.gov/region1/info/testmethods/pdfs/600490010_EPA_CompendMethod-

IndoorAirPollutants_P1.pdf . Published April 1990. Accessed May 2014.

8 NRC-Canada. Indoor Air Quality Guidelines and Standards, RR-204. http://archive.nrc-

cnrc.gc.ca/obj/irc/doc/pubs/rr/rr204/rr204.pdf. Published March 2005. Accessed May

2014.